HubSpot Builds Shared AI Foundation to Accelerate Product Delivery

According to HubSpot's blog post titled "How we Build with AI," the company outlines a three-phase transformation that began with productivity tooling and evolved into a shared AI platform. The post reports HubSpot rolled out a coding copilot in 2023 and saw 30% adoption in early stages. The post states HubSpot now has 100% of engineers using AI and reports a 73% increase in lines of code written by engineers. The piece is labeled as part one of a three-part series, with later parts covering Agent-first go-to-market and operating as an AI-first company.

Per the blog, HubSpot moved from using off-the-shelf copilots to building an internal shared foundation so multiple product surfaces could reuse capabilities and deliver a consistent customer experience. The post highlights work on answer engine optimization (AEO) for marketing teams and guidance for appearing in chat results, and it summarizes practical, GTM-facing tactics alongside engineering changes.

Companies that scale AI beyond point tools commonly invest in a shared layer that centralizes model access, prompt logic, and retrieval, reducing duplicated engineering effort. For practitioners, a shared foundation typically raises questions about observability, prompt/version management, and data governance across services. Those are common technical trade-offs documented in comparable engineering write-ups from other SaaS vendors.

Editorial analysis: HubSpot's writeup combines operational metrics with product guidance for marketers, which makes the post both an internal case study and a how-to for GTM teams focused on conversational discovery. For practitioners, the most actionable parts are the AEO recommendations and the documented adoption metrics, which help benchmark internal AI rollouts.

• •Adoption signals and quantitative metrics (engineering productivity, code volume) that other companies publish when they describe platform investments.
• •How HubSpot describes governance, telemetry, and model/versioning in future posts in this series.
• •The AEO tactics and whether they change indexing or content strategy for organizations that rely on search and conversational discovery.